Broadband near-field beam training method, device and medium

Abstract

The invention provides a broadband near-field beam training method and device and a medium, and belongs to the technical field of Internet communication transmission, and the method comprises the steps: obtaining a system model of a broadband super-large-scale array communication system; determining a first beam splitting effect of a sparse array in a broadband in a spatial domain and a second beam splitting effect of the broadband in a frequency domain based on the system model; and carrying out super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect. According to the technical scheme of the invention, super-resolution ultra-low-overhead broadband beam training can be realized.

Description

Technical Field

[0001] The present application relates to the field of Internet communication transmission technology, and particularly to a broadband near-field beam training method, device and medium. Background Art

[0002] In related technologies, narrowband beam training only uses one subcarrier on one time slot for beam training, and whether it is hierarchical beam training or multi-beam training, the beam training overhead is still proportional to the number of antennas in the array. For communication scenarios of ultra-large-scale arrays, the beam alignment delay is still unacceptable.

[0003] In addition, although broadband beam training can already control the beams of subcarriers to cover multiple angles and distances simultaneously, so as to achieve initial beam alignment with a small amount of pilot overhead, however, these broadband beam training methods can only ensure that the array gain loss in terms of angle and distance does not exceed 3 dB, and cannot achieve ultra-high-resolution beam alignment. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose a broadband near-field beam training method, device and medium, aiming to achieve ultra-high-resolution and extremely low-overhead broadband beam training.

[0005] To achieve the above object, a first aspect of the embodiments of the present application proposes a broadband near-field beam training method, and the method includes:

[0006] Obtain a system model of a broadband ultra-large-scale array communication system;

[0007] Based on the system model, determine a first beam splitting effect of a sparse array in the spatial domain and a second beam splitting effect of the broadband in the frequency domain;

[0008] Perform ultra-high-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0009] In some embodiments, the determining a first beam splitting effect of a sparse array in the spatial domain and a second beam splitting effect of the broadband in the frequency domain based on the system model includes:

[0010] Perform sparse activation on the system model to obtain a far-field channel model of a sparse linear uniform array;

[0011] Based on the far-field channel model, determine a first beam splitting effect of the sparse array in the spatial domain and a second beam splitting effect of the broadband in the frequency domain.

[0012] In some embodiments, the super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect includes:

[0013] Performing super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles;

[0014] Determining the actual user angle among the candidate user angles based on a first target subcarrier in the broadband; wherein the first target subcarrier meets a preset frequency interval requirement, and a single beam of the first target subcarrier covers the candidate user angles;

[0015] Performing distance estimation on the actual user angle based on the second beam splitting effect to obtain the user distance.

[0016] In some embodiments, the performing super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles includes:

[0017] In the case of activating a central sparse linear uniform array, based on the first beam splitting effect, using multiple rainbow blocks to perform super-resolution angle scanning to obtain multiple candidate user angles;

[0018] Wherein, the activated central sparse linear uniform array includes multiple antennas with equal antenna spacing; the interval between every two of the multiple rainbow blocks is less than zero, and the left edge of the first rainbow block among the multiple rainbow blocks covers the spatial angle -1, and the right edge of the last rainbow block covers the spatial angle 1.

[0019] In some embodiments, the determining the actual user angle among the candidate user angles based on a first target subcarrier in the broadband includes:

[0020] Obtaining the calibrated received power corresponding to each target subcarrier in the broadband;

[0021] Comparing the magnitudes of the calibrated received powers to obtain the highest calibrated received power among the calibrated received powers;

[0022] Determining a second target subcarrier corresponding to the highest calibrated received power among the target subcarriers;

[0023] Determining the candidate user angle covered by a single beam of the second target subcarrier among the candidate user angles as the actual user angle.

[0024] In some embodiments, the performing distance estimation on the actual user angle based on the second beam splitting effect to obtain the user distance includes:

[0025] When all antennas of the ultra-large-scale array are activated, the beam focusing of all subcarriers in the broadband is controlled by using the second beam splitting effect to focus on a specific position in the actual user angle, and the user distance is obtained.

[0026] Among them, all antennas of the ultra-large-scale array are activated based on a pilot.

[0027] To achieve the above object, a second aspect of the embodiments of the present application proposes a broadband near-field beam training device, and the device includes:

[0028] An acquisition module, configured to acquire a system model of a broadband ultra-large-scale array communication system;

[0029] A dual beam splitting effect determination module, configured to determine a first beam splitting effect of a sparse array in the broadband in the spatial domain and a second beam splitting effect of the broadband in the frequency domain based on the system model;

[0030] A beam training module, configured to perform super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0031] To achieve the above object, a third aspect of the embodiments of the present application proposes a computer device, and the computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect above is implemented.

[0032] To achieve the above object, a fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, and the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0033] To achieve the above object, a fifth aspect of the embodiments of the present application proposes a computer program product, and the computer program product stores a computer program. When the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0034] The broadband near-field beam training method, device, computer device, computer-readable storage medium, and computer program product proposed by the embodiments of the present application obtain a system model of a broadband ultra-large-scale array communication system, then determine a first beam splitting effect of a sparse array in the broadband in the spatial domain and a second beam splitting effect of the broadband in the frequency domain based on the system model, and finally perform super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0035] Thus, by simultaneously utilizing the dual beam splitting phenomena in the spatial domain and frequency domain based on the sparse array, the embodiment of the present application can achieve super-resolution beam alignment using limited spectral resources, not only improving the accuracy of beam training, but also greatly reducing the pilot overhead required for beam training, thereby realizing ultra-low overhead broadband beam training with super-resolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic flowchart of the steps of the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0037] Figure 2 It is a schematic diagram of a broadband very large-scale array downlink communication system involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0038] Figure 3 It is a schematic diagram of an activated S-ULA involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0039] Figure 4 It is a simulation diagram of the number of beams under different TD parameters involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0040] Figure 5 It is a schematic diagram of a rainbow block involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0041] Figure 6 It is a schematic diagram of the beam distribution within a rainbow block involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0042] Figure 7 It is a schematic diagram of the beam distribution at the edge of a rainbow block involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0043] Figure 8 It is a schematic diagram of the broadband beam training algorithm framework involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0044] Figure 9 For Figure 1 It is a schematic flowchart of the refined steps of step S103 in

[0045] Figure 10 It is a schematic diagram of angle ambiguity elimination in the case of the original beam coverage involved in the broadband near-field beam training method provided by the embodiment of the present application in some embodiments;

[0046] Figure 11Schematic diagram of angle ambiguity elimination in the case of calibration beam coverage involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0047] Figure 12 Schematic diagram of distance beam coverage involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0048] Figure 13 Graph of the change of the angle of angle estimation with the reference SNR involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0049] Figure 14 Graph of the change of the angle of distance estimation with the reference SNR involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0050] Figure 15 Graph of the change of the achievable rate with the reference SNR involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0051] Figure 16 Graph of the change of the achievable rate with the distance involved in the broadband near-field beam training method provided by the embodiments of the present application in some embodiments;

[0052] Figure 17 Schematic diagram of the structure of the broadband near-field beam training device provided by the embodiments of the present application;

[0053] Figure 18 Schematic diagram of the hardware structure of the computer device provided by the embodiments of the present application. Detailed implementation manners

[0054] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0057] First, a brief analysis of several technical terms involved in this application is given:

[0058] Millimeter wave (mmWave) and Terahertz (THz) communications are regarded as key technologies for the sixth-generation mobile communication standard (6G) wireless network due to their large bandwidths, which can significantly improve the capacity and spectral efficiency of wireless communication systems. In addition, since the wavelengths in the high-frequency band are shorter, more and more antennas can be packaged in a small area, so it is expected that extremely large-scale arrays (XL-arrays) will be possible in future wireless communications, thus significantly improving spectral efficiency and spatial resolution. However, the high-frequency band and XL-arrays also face two major challenges.

[0059] First, XL-arrays fundamentally change the radio propagation model, from a far-field plane wavefront to a near-field spherical wavefront (SW), which brings both benefits and many challenges. On the one hand, the beam focusing phenomenon greatly reduces the inter-user interference (IUI), thus improving the channel capacity. In addition, the near-field multiple-input multiple-output (MIMO) channel exhibits a high channel rank and achieves a high multiplexing gain, even in the line-of-sight (LoS) scenario. On the other hand, for beam training, near-field communication requires codewords in both the angular domain and the distance domain, which will result in a serious beam training overhead.

[0060] Secondly, since the beamforming structure based on phase shifters (PS) is frequency-independent, the extremely large bandwidth in the high-frequency band will introduce the so-called beam splitting effect, and the beams generated by different subcarriers (or frequencies) will be focused at different positions and deviate from the target position, resulting in a significant performance degradation.

[0061] For near-field beam training, related technologies have proposed a polarization-domain codebook, in which the angular domain and the range domain are sampled uniformly and non-uniformly, respectively. Although this method can effectively solve the energy diffusion problem when the traditional discrete Fourier transform (DFT) codebook is applied to near-field communication, it has an unaffordable beam training overhead, which is proportional to the product of the number of antennas and the sampling range. To solve this problem, a two-stage near-field beam training scheme has been proposed in related technologies, which decouples the angle and range estimation. Utilizing the energy diffusion effect in the received beam pattern, the user angle is estimated through the intermediate angle of the support angle region, and the user's distance is estimated through the above-mentioned polarization-domain codebook. In addition, to break the resolution limit, there is another distance off-grid beam training method, which utilizes the distance information behind the energy diffusion beam pattern. It has a similar beam training overhead to the two-stage near-field beam training method, but significantly improves the accuracy of distance estimation. However, the beam training overhead of the above two-stage near-field beam training method and the distance off-grid beam training method is still proportional to the number of antennas, which is too high in the XL-array scenario with a large number of antennas. To solve this problem, a hierarchical beam training scheme has been proposed in related technologies. First, a wide beam is used to estimate the rough user angle, and then narrow beams are used to gradually obtain more refined user distance and angle round by round. In addition, by using a novel antenna activation method to extend the multi-beam training scheme in traditional far-field communication to the near-field, problems such as coverage holes that occur when the subarray-based multi-beam generation method is applied to near-field communication can also be solved. In addition, related technologies have also proposed to use deep learning technology to reduce the overhead of near-field beam training. For this purpose, a deep neural network (DNN) is trained using traditional far-field codebooks and near-field codebooks, respectively.

[0062] However, the above methods are all narrowband beam training methods. Each time slot of the narrowband beam training method only utilizes a single subcarrier, which saves spectrum resources but increases the consumption of time resources. Therefore, in order to further reduce the beam training overhead, the related technology has proposed a broadband near-field beam training method, which essentially uses multiple subcarriers to cover different positions / directions in a pilot symbol. Thanks to the rich spectrum resources in the high-frequency band, broadband beam training can estimate the user's angle and range within several pilots. Specifically, a real-time delay (TD) beamforming structure can be used to control the beam splitting effect in near-field broadband communication and control the beams formed on different subcarriers to sequentially cover a specific distance ring, and its beam training overhead is the size of distance sampling. However, the overhead of this method still depends on the size of the array, and its beam training overhead is relatively high. To solve this problem, a new near-field broadband near-field beam training method based on distance-dependent beam splitting effect can be adopted, which controls the beams generated by different subcarriers to cover multiple inclined bands. Therefore, this method can cover multiple angles and distance rings simultaneously within one pilot, so that the beam training overhead can be lower. However, the above two broadband near-field beam training methods can only ensure that the loss does not exceed 3 dB compared with the array gain of the optimal beam, but cannot overcome the limited resolution limitation of angle and distance estimation under limited spectrum resources.

[0063] Next, the overall concept of the embodiments of the present application will be described.

[0064] For the narrowband beam training in the related technology, it only uses one subcarrier for beam training in one time slot. Whether it is hierarchical beam training or multi-beam training, its beam training overhead is still proportional to the number of antennas in the array. For the communication scenario of a very large-scale array, the delay of beam alignment is still unacceptable.

[0065] For the broadband beam training in the related technology, it has the following limitations:

[0066] First, it can only control the beam of the subcarrier to cover a specific distance ring. Its beam training overhead depends on the number of distance samplings, and the number of distance samplings is still positively correlated with the number of antennas in the very large-scale array. The beam training overhead is still very large. In addition, since the distance sampling can only ensure that the array gain does not lose 3 dB, the resolution of distance estimation is very low and cannot be guaranteed.

[0067] Second, although relying on the strip beam splitting phenomenon depending on the distance distribution, it can control the beams of the subcarriers to cover multiple angles and distances simultaneously, and realize the initial beam alignment with a small amount of pilot overhead. However, this method can only ensure that the array gain of the angle and distance does not lose 3 dB, and cannot achieve ultra-high-resolution beam alignment, and fails to effectively utilize the given spectrum resources.

[0068] Based on this, embodiments of the present application propose a broadband near-field beam training method, apparatus, computer device, computer-readable storage medium, and computer program product. By obtaining the system model of a broadband ultra-large-scale array communication system, and then determining the first beam splitting effect of a sparse array in the spatial domain and the second beam splitting effect of the broadband in the frequency domain based on the system model, and finally performing super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0069] In this way, embodiments of the present application utilize the dual beam splitting phenomena in both the spatial domain and the frequency domain based on a sparse array, and can complete super-resolution beam alignment by using limited spectrum resources, greatly improving the accuracy of beam training. Moreover, embodiments of the present application decouple angle and distance estimation by using the method of activating the central subarray, greatly reducing the pilot overhead required for beam training, and thus realizing a super-resolution and extremely low-overhead broadband near-field beam training method.

[0070] Based on the overall concept of the above embodiments of the present application, specific embodiments of the broadband near-field beam training method, apparatus, computer device, computer-readable storage medium, and computer program product provided by the embodiments of the present application are proposed. First, various specific embodiments of the broadband near-field beam training method in the embodiments of the present application are described in detail.

[0071] It should be noted that embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology, and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0072] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics technology, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0073] In addition, in each specific embodiment of the present application, when it comes to performing relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of such data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or redirecting to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0074] Furthermore, the broadband near-field beam training method provided by the embodiments of the present application relates to the field of Internet communication transmission technology. The broadband near-field beam training method provided by the embodiments of the present application can be applied to terminals, can also be applied to the server side, or can be software running on the terminal or the server side. In some embodiments, the terminal can be a base station (BS), a control and management device of the base station, a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, can also be configured as a server cluster or a distributed system composed of multiple physical servers, or can be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the broadband near-field beam training method, etc., but is not limited to the above forms.

[0075] Alternatively, the broadband near-field beam training method provided by the embodiments of the present application can also be used in many general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet-type devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, small computers, large computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0076] For the convenience of understanding and description, in the following text, the implementation of the broadband near-field beam training method provided by the embodiments of the present application in a terminal device will be taken as an example for detailed description. The implementation of the broadband near-field beam training method provided by the embodiments of the present application for any of the above forms of the subject can refer to the process of the broadband near-field beam training method implemented in the terminal device described later.

[0077] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of the steps of the broadband near-field beam training method provided by the embodiments of the present application in some embodiments. It should be understood that although Figure 1 shows the execution order of some method steps, based on different design requirements of actual applications, the broadband near-field beam training method provided by the embodiments of the present application can of course adopt an execution order different from that shown in the figure. That is, Figure 1 the order of the method steps shown does not constitute a limitation on the execution logic order of the broadband near-field beam training method provided by the embodiments of the present application, and any reasonable change based on Figure 1 the order of the method steps shown should be included in the protection scope of the broadband near-field beam training method provided by the embodiments of the present application.

[0078] As Figure 1 shown, in some embodiments, the broadband near-field beam training method provided by the embodiments of the present application may include but is not limited to steps S101 to S103.

[0079] Step S101: Obtain the system model of the broadband very large-scale array communication system.

[0080] Before the terminal device formally performs broadband beam training, it first performs a system modeling operation on the broadband very large-scale array communication system, so as to obtain the system model of the broadband very large-scale array communication system.

[0081] Exemplarily, the terminal device can perform modeling on the broadband very large-scale array downlink communication system as Figure 2 shown, so as to obtain the corresponding system model. As Figure 2 shown, in the broadband very large-scale array downlink communication system, the base station is equipped with a dense uniform linear array (D-ULA), which has N antennas (for convenience, it can be assumed that N is odd) and serves 1 single-antenna user.

[0082] In some embodiments, the system modeling operation of the terminal device on the broadband very large-scale array communication system includes channel modeling and beamforming design.

[0083] Among them, when the terminal device performs channel modeling, it is assumed that the D-ULA is placed along the y-axis and centered at the origin. Here, the nth antenna of the D-ULA is located at (0, nd c ), where represents the antenna index. The antenna spacing of the D-ULA is where represents the wavelength of the central subcarrier, and c and f c represent the speed of light and the frequency of the central subcarrier, respectively. The single-antenna user is located at where r 0 and θ 0 ∈[-1,1) represent the position and spatial angle from the base station to the user, respectively. It is assumed that the bandwidth is represented by B and is divided into M subcarriers. Specifically, the frequency of the mth subcarrier is given by the following formula:

[0084]

[0085] where represents the set of subcarrier indices.

[0086] For the very large-scale array scenario, it is assumed that the user is located in its Fresnel near-field region. Specifically, the distance from the base station to the user satisfies Z F <r 0 <Z Eff , where Z F = max{d R , 1.2D} and represent the Fresnel distance and the effective Rayleigh distance, respectively, and D = (N - 1)d c represents the array aperture. Therefore, the channel between the base station and the user can be modeled by a uniform spherical wavefront (USW). In addition, due to the severe path loss and shadow effects in high-frequency bands such as millimeter waves and terahertz waves, the line-of-sight path is considered, and the near-field channel from the base station to the user can be modeled as:

[0087]

[0088] where the parameter represents the line-of-sight path gain of the mth subcarrier, and λ m represents the wavelength of the mth subcarrier. Specifically, b m (r 0 , θ 0 ) represents the response vector of the near-field channel, specifically as follows:

[0089]

[0090] where represents the distance between the nth antenna and the user. Since r n is a complex root function and is difficult to analyze. To solve this problem, the Fresnel approximation can be utilized, which has been proven to be accurate in the Fresnel region. Then, r n can be approximated as:

[0091]

[0092] where,

[0093] In addition, when the terminal device performs beamforming design, since the channel in the near-field channel model shown in the above formula (1) is frequency-dependent, when only applying phase shifter (PS)-based beamforming that is independent of frequency, the beam splitting effect may significantly degrade the rate performance of the broadband communication system. To solve this problem, beamforming based on real-time time delay TD is considered. It forms frequency-dependent beams and thus becomes an effective method to compensate for or control the near-field beam splitting effect. Specifically, the very large-scale array is connected with N TD circuits and N PSs, where each antenna is connected to a real-time time delay TD circuit and a phase shifter PS. Each TD circuit can adjust the frequency-dependent phase shift by introducing a controllable delay on the broadband signal. Mathematically, the TD beamformer of the very large-scale array is represented by and can be expressed as:

[0094]

[0095] where, τ n represents the adjustable delay of the nth TD circuit. Specifically, to match the near-field channel, can be set, where and μ′ represent adjustable TD parameters. Then, the TD beamformer can be rewritten as:

[0096]

[0097] This is similar to the form of the channel response vector in the above formula (2). Similar to TD beamforming, the PS beamformer of the very large-scale array is given by:

[0098]

[0099] where, θ′ p and μ′ p represent the PS angle and range parameters, respectively. Based on the above, the effective beamformer w m (θ′, μ′, θ′ p , μ′ p ) of the very large-scale array can be written as:

[0100] w m (θ′, ; μ′; θ′ p , μ′ p ) = w m (θ′, μ′)☉w PS (θ′ p , μ′ p ). (6)

[0101] Then, the signal received by the user at the m-th subcarrier is given by:

[0102]

[0103] where P t and x m represent the transmit power and the signal, where In addition, represents the additive white Gaussian noise (AWGN), where σ 2 represents the noise power.

[0104] Step S102: Determine the first beam splitting effect of the sparse array in the spatial domain and the second beam splitting effect of the broadband in the frequency domain based on the system model.

[0105] After the terminal device obtains the system model of the broadband very large scale array communication system, it further performs dual beam splitting control of the sparse array in the spatial domain and the frequency domain based on this system model, so as to determine the first beam splitting effect of the sparse array in the spatial domain of the broadband and the second beam splitting effect of the broadband in the frequency domain.

[0106] In some embodiments, by describing the beam characteristics of the sparse array in the broadband, the terminal device can obtain results indicating that both the sparse array and the broadband cause beam splitting effects, corresponding to the spatial domain and the frequency domain respectively.

[0107] In some embodiments, the above step S102 may include:

[0108] Perform sparse activation on the system model to obtain the far-field channel model of the sparse linear uniform array;

[0109] Determine the first beam splitting effect of the sparse array in the spatial domain and the second beam splitting effect of the broadband in the frequency domain based on the far-field channel model.

[0110] The terminal device can describe the broadband multi-beam characteristics through the far-field channel model of the sparse linear uniform array S-ULA with sparse activation, so as to determine the first beam splitting effect of the sparse array in the spatial domain and the second beam splitting effect of the broadband in the frequency domain.

[0111] In some embodiments, the activated S-ULA obtained by the terminal device through sparse activation of the system model is as Figure 3 shown. When the terminal device acquires the far-field channel model of the sparsely activated S-ULA, it first selects a central subarray with Q tol antennas. Assuming that the user is located in the far-field region of the subarray, that is Then, uniformly activate antennas of the central subarray, and deactivate (U - 1) antennas in the middle, as Figure 3 shown. Specifically, there is assuming it is an integer. Therefore, the central subarray with Q tol antennas becomes an S-ULA with antennas. Therefore, the line-of-sight LoS channel between the S-ULA and the user at the m-th subcarrier can be modeled under the traditional plane wavefront as follows:

[0112]

[0113] where, a m (θ 0 , U) represents the far-field array response vector of the activated S-ULA. Mathematically, a m (θ 0 , U) can be expressed as:

[0114]

[0115] where, represents the antenna index set of the S-ULA. Among them, the far-field TD beamformer of the S-ULA is given by the following formula:

[0116]

[0117] where, represents the adjustable TD parameter, and the actual delay of the -th TD circuit in the activated S-ULA can be expressed as

[0118] When the terminal device describes the broadband multi-beam characteristics of the sparsely activated S-ULA, let f m (θ, θ′ SA , U) represent the array gain of the activated central S-ULA at the observation angle θ, and its TD beamformer is set to w m (θ′ SA , U). Therefore, f m (θ, θ′ SA , U) can be given by the following formula:

[0119]

[0120] Based on Equation (11), the following results can be obtained:

[0121] Lemma 1: Consider an active center S-ULA parameterized by U and a TD beamformer w m (θ′ SA , U). beams will be formed on the m-th subcarrier, where each beam angle is given by:

[0122]

[0123] where, and where represents the set of integers. Without loss of generality, it is usually assumed that Thus, it can be obtained that

[0124] Proof: The array gain in the above Equation (11) can be further simplified as:

[0125]

[0126] where, (a 1 ) can be determined from the reference "Zhou C, You C, Huang Z, et al. Multi-beam training for near-field communications in high-frequency bands[J]. arXiv preprint arXiv:2406.14931, 2024". By setting that is where, it can be obtained that f(θ, U) = Q. Considering that is the actual spatial angle that satisfies should be constrained as:

[0127]

[0128] Therefore, beams are formed on the m-th subcarrier in the angle domain. Q.E.D.

[0129] From Lemma 1, it can be seen that the period of the multi-beam formed at the m-th subcarrier is Therefore, or depending on the adjustable TD parameter θ′ SA. Specifically, the beam period at the central sub - carrier is so as to form 2 / (2 / U) = U beams in the angular domain. Generally, the number of multi - beam depends on the TD parameter θ′ SA and the sub - carrier frequency f m , which will be further elaborated below.

[0130] Lemma 2: Consider an S - ULA with parameter U and a TD beamformer w m (θ′ SA , U). The number of split beams formed at the m - th sub - carrier is given by:

[0131]

[0132] Proof: For the case of, it can be obtained that:

[0133]

[0134] This shows that beams are generated on the m - th sub - carrier. For the other case there is:

[0135]

[0136] This means that only beams are generated, thus completing the proof.

[0137] Based on Lemma 2 and the condition B << f c (i.e., ρ m ≈ 1), there is:

[0138]

[0139] Therefore, for sub - carriers with f m > f c , the number of split beams is U or U + 1, while for sub - carriers with f m < f c , the number of split beams is (U - 1) or U. Generally, in the angular domain, MU beams are simultaneously generated on M sub - carriers. By effectively using these beams for angular scanning, a higher angular resolution can be achieved compared to a traditional half - wavelength - spaced array with the same frequency resources.

[0140] Exemplarily, Example 1: In Figure 4In the beam number simulation diagrams under different TD parameters shown, the terminal device plots the relationship between the array gain at the m-th frequency and the spatial angle. Assume that a central S-ULA activates only Q = 17 antennas with an activation interval of U = 8. Assume that the base station BS operates at a central frequency f c = 60 GHz with a bandwidth of B = 3 GHz and has M = 1024 subcarriers. Consider multi-beams at a frequency of f H = f M = 61.5 GHz.

[0141] For the TD parameter θ′ SA = -1.4596 and there are:

[0142]

[0143] Therefore, it can be observed in Figure 4 the number of beams.

[0144] For the TD parameter θ′ SA = -0.9021 and there are:

[0145]

[0146] Therefore, the number of Figure 4 beams are formed on the M-th subcarrier, as shown in

[0147] It should be noted that the above Example 1 verifies the multi-beam characteristic analysis in the above formula (14).

[0148] In some embodiments, in order for the terminal device to obtain controllable beam splitting, the beam patterns of all subcarriers are mainly discussed in the following two cases: Case 1) θ′ SA ∈[-1,1); Case 2)

[0149] Lemma 3: When the TD parameter satisfies θ′ SA ∈[-1,1), there is one beam in each subcarrier facing the same angle

[0150] Proof: When there are: This is a physical angle, resulting in Therefore, one beam is formed on each subcarrier, and its direction points to the same angle Proof completed.

[0151] Lemma 3 shows that: When θ′ SAWhen ∈[-1,1), M beams are not effectively utilized to cover different angles, which is not the desired scheme in subsequent beam training design. Therefore, this situation is called an invalid TD parameter. Therefore, the main discussion is about the case of. In this case, the beam coverage of the rainbow block can be controlled to achieve super-resolution angle estimation. The following mainly focuses on θ′ SA <-1, and a similar analysis can be obtained for the case of θ′ SA >1. For θ′ SA <-1, there is:

[0152]

[0153] Then, under the condition of θ′ SA <-1, the multi-beam pattern of all subcarriers is analyzed in detail.

[0154] First, a rainbow block is defined as follows.

[0155] Definition 1: (Rainbow block). For each subcarrier, one of the multi-beam angles that shares the same with the central subcarrier is collected into a set. This set is called the rainbow block with respect to the parameter . Specifically, since there are U rainbow blocks. Mathematically, the u c -th rainbow block is defined as:

[0156]

[0157] where

[0158] For this purpose, the angular interval is defined as the coverage area of the u c -th rainbow block. It should be noted that for the first and last rainbow blocks, there may be subcarriers such that:

[0159]

[0160] In other words, the actual angle where while the actual angle does not exist. However, this situation will not affect the subsequent beam training scheme design. Without loss of generality, it is chosen to reuse to represent the angle regardless of whether it physically exists or not. To analyze the beam characteristics of the rainbow block, several definitions are given below.

[0161] Definition 2: (Width of the rainbow block). Given the TD parameter θ′ SA , the width of the u c -th rainbow block is defined by the coverage area of the u c -th rainbow block. Mathematically, there is:

[0162]

[0163] Definition 3: (Interval between rainbow blocks). Given the TD parameter θ′ SA , the rainbow block interval between the u c -th rainbow block and the (u c + 1)-th rainbow block is defined as:

[0164]

[0165] According to the above definitions, it is found that the beam pattern of all subcarriers is mainly composed of U rainbow blocks. In addition, the width of the rainbow block increases with the increase of the rainbow block index u c , as shown in Figure 5 . In addition, it can be seen from formula (19) that since -B < 0, the rainbow interval decreases with the decrease of . When the rainbow interval , this situation is called achieving seamless beam coverage in the angular region of . Next, discuss how to use the beams in U rainbow blocks to seamlessly cover the entire angular domain.

[0166] First, let the central subcarrier direct U beams towards the angle to obtain the following constraints:

[0167]

[0168] Then, impose the following two conditions to seamlessly cover the entire angular domain [-1, 1).

[0169] Condition 1: The interval between every two rainbow blocks is less than zero, that is

[0170] Condition 2: The left edge of the first rainbow block covers -1, while the right edge of the last rainbow block covers 1.

[0171] It should be noted that Condition 1 means that the angular range is seamlessly covered by the beams within U rainbow blocks, and Condition 2 further ensures that the entire angular domain [-1, 1) is covered. Since the rainbow block interval between the first and the second rainbow blocks is relatively wide, if RG (1) ≤ 0, then there is: Due to the rainbow block interval As increases, Condition 1 leads to the following constraint:

[0172] (Constraint 1) RG (1) ≤0. (21)

[0173] Then, Condition 2 has the following constraint:

[0174]

[0175] In this way, seamless beam coverage can be achieved within the angular domain [-1, 1). Next, a feasible TD parameter θ′ SA is given as follows.

[0176] Lemma 4: Given the constraints (21)-(23), a feasible TD parameter θ′ SA can be set as:

[0177]

[0178] where,

[0179] Proof: Based on (20), by substituting into the constraints (22) and (23), we can obtain:

[0180]

[0181] Then, the constraint (23) can be simplified to:

[0182]

[0183] Combining equations (25) and (26), we can get:

[0184]

[0185] Substituting equation (27) into equation (20), a feasible solution for the TD parameter θ′ SA can be obtained, that is: Thus, the proof is completed.

[0186] Based on Lemma 4, seamless beam coverage in the angular domain can be achieved by setting . However, the coverage areas of each adjacent rainbow block will overlap slightly, resulting in denser coverage at the boundaries. In fact, the TD parameter value selected in equation (24) is the solution with the smallest overlapping coverage area between the rainbow blocks. It can be observed that the interval between rainbows As increases with the increase of. Therefore, select (see the conditions in (27)), thus generating a smaller overlapping area In addition, some beam angles in the overlapping coverage area may turn towards the same angle, that is where In this case, there is:

[0187] However, the angle coverage of adjacent beams within the overlapping area is very close, resulting in low utilization rate of multi-beam resources. Therefore, it is necessary to avoid too large an overlapping area between rainbow blocks. The maximum overlapping area is |RG (U-1) |, as follows:

[0188]

[0189] It can be observed that the maximum overlapping area is proportional to the relative bandwidth For the base station BS parameters in the above Example 1, there is: |RG (U-1) | = 0.1, which is much smaller than the width of the rainbow block. Since the overlapping area is relatively small, it can be roughly assumed that MU beams are evenly covered in the angular domain from -1 to 1.

[0190] As Figure 6 and Figure 7 shown, the multi-beam diagrams of all sub-carriers within and at the boundary of the rainbow block are respectively plotted in Figure 6 and Figure 7 . The base station BS parameters are the same as those in the above Example 1. Figure 6 shows that the beams within each rainbow block are evenly covered in a specific angular region, and approximately U = 8 beams are distributed within an angular range of every 0.02. In contrast, only one beam coverage can be generated in the same space in the related art, which indicates that the broadband near-field beam training method proposed in the embodiments of the present application based on a sparse array can significantly improve the resolution under the condition of the same frequency resources. In addition, it can be seen from Figure 7 that although the multi-beams of different sub-carriers are more densely covered at the boundary, the angles of many beams are very close, and the resolution cannot be improved even under high signal-to-noise ratio (SNR) conditions, resulting in low utilization efficiency of multi-beam resources.

[0191] Step S103: Perform super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0192] After the terminal device determines the double beam splitting effect in the spatial domain and frequency domain of the sparse array, it further utilizes this double beam splitting effect in the spatial domain and frequency domain (i.e., the above-mentioned first beam splitting effect and second beam splitting effect) to perform effective super-resolution broadband near-field beam training.

[0193] In some embodiments, as Figure 8 shown, the broadband beam training performed by the terminal device based on the double beam splitting effect consists of three stages, namely: angle scanning using multiple rainbow blocks, angle ambiguity elimination, and distance scanning.

[0194] It should be noted that the key idea of the broadband beam training performed by the terminal device based on the double beam splitting effect is to utilize the beam splitting effect of the activated S-ULA in the frequency domain and spatial domain to expand the beam coverage area and achieve super-resolution angle estimation. Then, utilize the characteristic that the single beams on each subcarrier are roughly evenly distributed in the angle domain to select specific subcarriers with appropriate frequency intervals. Use the single beams of these selected subcarriers to cover the candidate user angles, and solve the angle ambiguity by comparing the calibrated received powers on the selected subcarriers. Finally, by activating the entire XL-array, control the split beams on all subcarriers to focus on the estimated user angles but different distances, which only requires one pilot to achieve distance scanning.

[0195] In the embodiments of the present application, by the terminal device simultaneously utilizing the double beam splitting phenomenon based on the spatial domain and frequency domain of the sparse array, super-resolution beam alignment can be completed using limited spectral resources, greatly improving the accuracy of beam training. Moreover, in the embodiments of the present application, by the terminal device using the method of activating the central subarray, the angle and distance estimations are decoupled, greatly reducing the pilot overhead required for beam training, and thus realizing a super-resolution extremely low-overhead broadband near-field beam training method.

[0196] Please refer to Figure 9 , Figure 9 For Figure 1 the detailed step flow schematic diagram of step S103 in

[0197] As Figure 9 shown, in some embodiments, the above-mentioned step S103: performing super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect may include steps S901 to S903 as follows.

[0198] Step S901: Perform super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles.

[0199] When the terminal device performs broadband beam training based on the above-mentioned dual beam splitting effect, it first performs super-resolution angle estimation based on the first beam splitting effect of the activated central sparse linear uniform array (S-ULA) in the spatial domain, so as to obtain multiple candidate angles.

[0200] It should be noted that the terminal device can unify the two stages of angle scanning and ambiguity elimination into angle estimation. Specifically, due to sparsity, periodic multi-beams will generate angle ambiguities, which can be solved by activating the central sub-array. In addition, in the foregoing description, it has been revealed that the activated central S-ULA has a beam splitting effect (the first beam splitting effect) in the spatial domain, which can be used for super-resolution angle scanning. Therefore, the terminal device can first perform super-resolution angle estimation based on the first beam splitting effect of the activated central S-ULA in the spatial domain.

[0201] In some embodiments, step S901 above: performing super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles may include:

[0202] In the case of activating the central sparse linear uniform array, based on the first beam splitting effect, use multiple rainbow blocks to perform super-resolution angle scanning to obtain multiple candidate user angles.

[0203] It should be noted that the activated central sparse linear uniform array includes multiple antennas with equal antenna spacings. For example: the central S-ULA has Q antennas with an antenna spacing of Ud 0 In addition, to ensure seamless coverage of subcarriers in the broadband over the entire angular domain, the interval between every two rainbow blocks among the multiple rainbow blocks is less than zero, and the left edge of the first rainbow block among the multiple rainbow blocks covers the spatial angle -1, and the right edge of the last rainbow block covers the spatial angle 1.

[0204] Based on the fact that the first beam splitting effect of the activated central S-ULA in the spatial domain can be used for super-resolution angle scanning, the terminal device first activates a central S-ULA, which has Q antennas with an antenna spacing of Ud 0 The TD beamformer is given by the above formula (10), and the TD parameter θ' SA can be given by the above formula (24). Then, the received signal of the m-th subcarrier is given by the following formula:

[0205]

[0206] However, since the higher the frequency, the greater its path loss, a calibrated received power is defined on the m-th subcarrier, and its formula is:

[0207]

[0208] Therefore, it is possible to estimate the subcarriers with the highest calibrated received power. In addition, due to the beam splitting effect in the spatial domain, the angle of the k-th candidate user is given by:

[0209]

[0210] where and

[0211] Step S902: Determine the actual user angle among the candidate user angles based on the first target subcarriers in the wideband; wherein, a single beam of the first target subcarriers covers the candidate user angles.

[0212] After the terminal device obtains multiple candidate angles through angle scanning, it further selects a set of first target subcarriers with an appropriate frequency interval from multiple subcarriers in the wideband, and determines the actual user angle among the candidate user angles by using the corresponding single beam to cover the candidate user angles on the first target subcarriers.

[0213] Exemplarily, similar to the sparse activation method, the terminal device can activate the central subarray (dense array) composed of Q antennas to solve the angle ambiguity problem. Among them, it is assumed that the user is located in the far-field region of the activated central subarray. Specifically, select a set of subcarriers, and use the corresponding single beam (a total of beams) on each selected subcarrier to cover the candidate user angles. Then, according to the subcarrier with the highest calibrated received power selected, the actual user angle can be obtained.

[0214] In some embodiments, the above step S902: Determine the actual user angle among the candidate user angles based on the first target subcarriers in the wideband may include:

[0215] Obtain the calibrated received power corresponding to each target subcarrier in the wideband;

[0216] Compare the magnitudes of the calibrated received powers to obtain the highest calibrated received power among the calibrated received powers;

[0217] Determine the second target subcarrier corresponding to the highest calibrated received power among the target subcarriers;

[0218] Determine the candidate user angle covered by the single beam of the second target subcarrier among the candidate user angles as the actual user angle.

[0219] When determining the actual user angle among candidate user angles, the terminal device obtains the calibrated received power corresponding to each target subcarrier in the broadband, then compares the magnitudes of the calibrated received powers in sequence to obtain the highest calibrated received power among the calibrated received powers. After that, the terminal device further determines the second target subcarrier among the target subcarriers in the broadband that corresponds to the highest calibrated received power, and determines the candidate user angle covered by a single beam of the second target subcarrier among the candidate user angles as the actual user angle.

[0220] Exemplarily, the terminal device first gives the channel model of the central subarray and the corresponding TD beamformer. In this way, the LoS channel between users at the m-th subcarrier can be modeled as:

[0221]

[0222] where, a m (θ 0 ) represents the far-field array response vector that activates the central subarray, which is given by:

[0223]

[0224] where, represents the antenna index set of the central subarray. In addition, the TD beamformer is given by the following formula:

[0225]

[0226] where, represents the adjustable TD parameter, represents the actual time delay of the q-th TD circuit in the activated central subarray. Next, it is necessary to sequentially direct the single beams on the subcarriers to the candidate user angles. To achieve this goal, the array gain of the central subarray is first discussed, and a subcarrier selection method is proposed.

[0227] Similar to the above formula (11), the array gain of the m-th subcarrier is given by:

[0228]

[0229] Then, the angle of the beam generated by the m-th subcarrier can be obtained as follows.

[0230] Lemma 5: Given an activated central subarray with Q antennas and a TD beamformer w m (θ′ SA ), the steering beam angle θ m at frequency f m is given by:

[0231]

[0232] Among them,

[0233] Proof: The proof is similar to Lemma 1, and the same content will not be elaborated here.

[0234] According to Lemma 5, the beam period at subcarrier is This results in the subcarrier having a single-beam characteristic. On the contrary, for subcarrier the period is This may result in multiple beams and introduce new angle ambiguities.

[0235] Therefore, we limit the consideration range to subcarrier In addition, without loss of generality, an appropriate TD parameter θ′ CS , can be set so that the p m of all subcarriers remains the same, with the condition:

[0236]

[0237] where Then, the guiding beam angle at frequency is corrected to

[0238] Next, the subcarrier selection mechanism and the corresponding TD parameter values are introduced in detail. First, it is proved that the single beams on each subcarrier are approximately uniformly distributed in the angle domain. Then, this uniformity can be used to select subcarriers with appropriate frequency intervals, and use their single beams to cover the periodic candidate user angles, thus solving the angle ambiguity problem.

[0239] Lemma 6: Given the central subcarrier frequency f c and bandwidth B and M subcarriers, the angle difference between the single beams formed at adjacent subcarriers is given by:

[0240]

[0241] Proof: We can regard the single-beam angle in formula (36) as a function of frequency, as follows:

[0242]

[0243] Since B << f c , θ(f) can be obtained by expanding θ(f) at the central frequency f cApproximated by the first-order Taylor expansion as follows:

[0244]

[0245] Therefore, at the m L th subcarrier, the beam angle can be approximated as:

[0246]

[0247] Then, the difference in the guiding beam angles between adjacent subcarriers is Thus, we have completed the proof.

[0248] According to Lemma 6, subcarriers can be selected with a uniform spacing of ranging from f L to f c to obtain a beam period of ηΔθ. Specifically, the selected subcarriers are given by the following formula:

[0249]

[0250] To ensure that the single beam on the selected subcarriers is accurately aligned with the candidate user angle, the TD parameter θ′ CS should be set to:

[0251]

[0252] Lemma 7: Given the constraints (41) and (42), a feasible solution for the TD parameter θ′ CS is:

[0253]

[0254] Proof: Substituting into formula (42), we get:

[0255]

[0256] where Since p is an integer, a feasible solution for θ′ CS can be approximated as Thus, the proof is completed.

[0257] Due to the first-order Taylor approximation in formula (37), some of the single beams on the selected subcarriers may deviate slightly from the candidate user angle. Given the approximate TD parameter θ′ CS in formula (43), the subcarrier selection in formula (40) needs to be calibrated, and the calibrated subcarriers are given by the following formula:

[0258]

[0259] Among them,

[0260] Please refer to Figure 10 and Figure 11 , Figure 10 which is a schematic diagram of angle ambiguity cancellation in the case of the original beam coverage involved in the broadband near-field beam training method provided by the embodiments of this application in some embodiments. Figure 11 which is a schematic diagram of angle ambiguity cancellation in the case of the corrected beam coverage involved in the broadband near-field beam training method provided by the embodiments of this application in some embodiments.

[0261] In Figure 10 and Figure 11 the array gain of the selected subcarriers is plotted. The base station parameters are the same as those in Example 1 above. Assume that the best subcarrier obtained through angle scanning is f 300 = 59.3774 GHz, corresponding to the candidate user angles -0.8199165 + 0.2526k, k = 1, 2, L, 8. The subcarriers selected by formula (40) are f (k) = 30 - 0.1875k GHz. Figure 10 shows that the beams generated by the uncalibrated f (7) and f (8) deviate from the candidate user angles and which may reduce the performance of broadband beam training. However, the selected frequencies can be refined by formula (44), where f (6) , f (7) and f (8) are calibrated to 59.0698, 58.8853, and 58.7036 GHZ respectively. It can be seen from Figure 11 that the single beam on the calibrated subcarriers is precisely aligned with the candidate user angles, thus improving the angle estimation accuracy.

[0262] Given the TD parameter θ' CS in formula (43) and the selected subcarriers in formula (44), the received signal at the kth selected subcarrier is given by the following formula:

[0263]

[0264] Similar to (31), the subcarrier with the highest calibrated power selected can be estimated as:

[0265]

[0266] It corresponds to the estimated user angle (actual user angle), which is given by:

[0267]

[0268] In this way, in the subsequent distance scanning stage of the terminal device, it only needs to scan the distance domain within the estimated angle θ * to significantly reduce the near-field beam training overhead.

[0269] Step S903: Estimate the distance of the actual user angle based on the second beam splitting effect to obtain the user distance.

[0270] After the terminal device solves the angle ambiguity problem and determines the actual user angle among the candidate user angles, it further estimates the distance of the actual user angle based on the second beam splitting effect, and then obtains the user distance. For example, the terminal device activates the entire XL-array and then controls the split beams on all subcarriers to focus on the estimated user angle but different distances, which can achieve distance scanning with only one pilot.

[0271] In some embodiments, step S903: Estimate the distance of the actual user angle based on the second beam splitting effect to obtain the user distance, may include:

[0272] In the case of activating all antennas of the very large-scale array, use the second beam splitting effect to control the beams of all subcarriers in the broadband to focus on a specific position in the actual user angle to obtain the user distance.

[0273] It should be noted that all antennas of the very large-scale array are activated based on one pilot.

[0274] During the distance estimation stage of the terminal device, it can activate the antennas of the entire XL-array and control the beams on all subcarriers to focus on a specific position in the actual user angle determined in the angle estimation stage. Specifically, only one pilot is needed to achieve distance scanning using the broadband beam splitting effect, without the need for an exhaustive search in the range domain.

[0275] Exemplarily, the array gain of the entire XL-array is:

[0276]

[0277] Then, the beam focusing points of different subcarriers can be obtained in the following way.

[0278] Lemma 8: Given a TD-PS beamformer w m (θ′, μ′, θ′ p , μ′ p ), the beam focusing point (θ m , μ m ), μm ) can be expressed as:

[0279]

[0280] where and

[0281] Proof: The array gain of the m-th subcarrier in formula (47) can be expressed as:

[0282]

[0283] where In addition, it can be observed that F(x, y) is a periodic function with a period of

[0284] Mathematically, there is: Considering and max F(x, y) = 1, by setting and the focusing position can be obtained. Thus, results (48) and (49) are obtained. Q.E.D.

[0285] Lemma 8 shows that when setting θ′ = θ * and θ′ p = 0, the beams on all subcarriers will be focused at the same angle θ * which is the desired TD-PS angle parameter. In addition, considering and the beam period in the distance domain is Therefore, only one beam will be formed on the m-th subcarrier in the distance domain. Without loss of generality, it can be assumed that all subcarriers share the same s m = 0 for the TD parameter μ′, and it is necessary to satisfy Actually, it is necessary to control the subcarriers to be focused within a certain distance range [r min , r max , corresponding to the distance ring range [μ min , μ max , where To cover the distance range [r min , r max , first focus the beam of the central subcarrier on where Mathematically, there is:

[0286]

[0287] Therefore, to cover the required distance range, the following two conditions should be satisfied:

[0288]

[0289] For the TD-PS parameters μ′ and μ′ p A feasible solution that satisfies the above conditions is as follows.

[0290] Lemma 9: Given conditions (51) and (52), the TD-PS parameters μ′ and μ′ p A feasible solution is given by the following formula:[[]]

[0291]

[0292] Proof: Substitute into formulas (51) and (52), and we can get:[[]]

[0293]

[0294] Combining formulas (55) and (56), we can get:[[]]

[0295]

[0296] Therefore, the PS parameter μ′ p can be set as μ′ p = μ th . Then, the TD parameter μ′ is given by , Q.E.D.

[0297] Please refer to Figure 12 . Figure 12 This is a schematic diagram of distance beam coverage involved in the broadband near-field beam training method provided by the embodiments of this application in some embodiments.

[0298] In Figure 12 the beam patterns of some subcarriers in the distance domain are plotted. The number of antennas of the entire XL-array is set to N = 513. Other BS parameters are the same as those in Example 1. For clarity, only the beam at subcarrier f 1:30:1021 is plotted. The TD parameters are set to θ′ = 0 and μ′ = -1.7918, while the PS parameters are θ′ p = 0 and μ′ p = 1.8468. In Figure 12 it can be observed that the beams on the above M subcarriers are focused within the desired range [10, 50] m. In fact, the average distance resolution can reach where Δ r = r max - r min . However, since the subcarrier coverage density decreases with the increase of distance (as shown in Figure 12 ), this resolution cannot be achieved when users are evenly distributed.

[0299] Given the PS parameter (θ′p = 0, μ' p = μ th ) and TD parameters The received signal at the m-th subcarrier is expressed as:

[0300]

[0301] The subcarrier with the highest calibrated received power is given by the following formula:

[0302]

[0303] where the corresponding estimated user distance is:

[0304]

[0305] Next, a simulation analysis of the broadband training method provided by the embodiments of the present application is proposed.

[0306] The system parameters of the base station BS are set as follows: It is assumed that BS is equipped with N = 513 antennas, activates a central S-ULA, the activation interval U = 8, and the number of antennas of the central dense subarray activated in the second stage is Q = 65. The central carrier frequency is f c = 60 GHz, the bandwidth is B = 3 GHz, and there are M = 1024 subcarriers. The transmit power and noise power of BS are set to P t = 30 dBm and σ 2 = -80 dBm. In addition, the reference SNR is defined as while the normalized mean square error (NMSE) of angle and distance estimation are respectively defined as and Finally, all numerical results are obtained in 1000 channel realizations.

[0307] Please refer to Figure 13 in Figure 13The relationship between the angle estimation accuracy and the reference SNR is plotted, where the user angles and distances are randomly distributed within the ranges of [-1, 1] and [10m, 50m]. Several important observations can be summarized. First, the results show that the angle estimation NMSE of the broadband near-field beam training method provided by the embodiments of the present application (illustrated as "the proposed broadband super-resolution beam training method") decreases as the reference SNR increases and is significantly lower than all benchmark schemes. This is because the broadband near-field beam training method provided by the embodiments of the present application achieves higher resolution by forming more beams in the angle domain using the activated S-ULA, resulting in smaller angle estimation errors. Second, in the low SNR state, the angle estimation NMSE of the beam training scheme based on the near-field rainbow is close to that of the broadband near-field beam training method provided by the embodiments of the present application. This can be explained that the broadband near-field beam training method provided by the embodiments of the present application is generally more sensitive to noise, resulting in lower expected accuracy.

[0308] Please refer to Figure 14 , Figure 14 which shows the relationship between the distance estimation NMSE and the reference SNR. It can be observed that the distance estimation NMSE of the broadband near-field beam training method provided by the embodiments of the present application (illustrated as "the proposed broadband super-resolution beam training method") is significantly better than all benchmark schemes, especially under high SNR conditions. This can be explained that the benchmark schemes are on-grid beam training methods and use predefined distance codewords for distance estimation, which depends on the number of distance samples V, resulting in lower accuracy. However, the broadband near-field beam training method provided by the embodiments of the present application uses a large number of beams formed on different subcarriers to focus on the desired angles, making the distance estimation accuracy depend on the number of subcarriers. Therefore, due to M >> L, the proposed broadband beam training achieves better resolution / accuracy.

[0309] In addition, please refer to Figure 15 , Figure 15 where the achievable rate of different beam training schemes versus the reference SNR is plotted. First, it can be observed that the broadband near-field beam training method provided by the embodiments of the present application (illustrated as "the proposed broadband super-resolution beam training method") is superior to other benchmark schemes and is close to the performance of beamforming based on perfect CSI. This is because the broadband near-field beam training method provided by the embodiments of the present application achieves super-resolution estimation in both the angle and range domains. In addition, in the low SNR state, the performance of the two-stage beam training is significantly worse than other methods. This is because the two-stage beam training uses the energy diffusion effect to estimate the user angle, which has lower angle estimation accuracy in the low SNR state.

[0310] Finally, please refer to Figure 16 , Figure 16Shows the relationship between the achievable rate and the user distance. The users are uniformly distributed in the angular sector θ ∈ [-1, 1]. The following are the observations: Observation 1, the achievable rate performance of all schemes decreases as the user distance increases. This is because the reference SNR decreases as the user range increases, thus reducing the achievable rate performance of all schemes. Observation 2, for all user distances, due to the sparse activation method achieving higher resolution, the broadband near-field beam training method provided by the embodiments of the present application (illustrated as "the proposed broadband super-resolution beam training method") is superior to all other near-field beam training schemes.

[0311] Based on the same technical concept as the above broadband near-field beam training method, the embodiments of the present application further provide a broadband near-field beam training device that can implement the above broadband near-field beam training method.

[0312] Please refer to Figure 17 , the broadband near-field beam training device provided by the embodiments of the present application includes:

[0313] An acquisition module, configured to acquire the system model of a broadband massive MIMO communication system;

[0314] A dual beam splitting effect determination module, configured to determine a first beam splitting effect of a sparse array in the spatial domain and a second beam splitting effect of the broadband in the frequency domain based on the system model;

[0315] A beam training module, configured to perform super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

[0316] In some embodiments, the dual beam splitting effect determination module is further configured to perform sparse activation on the system model to obtain a far-field channel model of a sparse linear uniform array; and, determine a first beam splitting effect of a sparse array in the spatial domain and a second beam splitting effect of the broadband in the frequency domain based on the far-field channel model.

[0317] In some embodiments, the beam training module is further configured to perform super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles; determine the actual user angle among the candidate user angles based on a first target subcarrier in the broadband; wherein, a single beam of the first target subcarrier covers the candidate user angles; and, perform distance estimation on the actual user angle based on the second beam splitting effect to obtain the user distance.

[0318] In some embodiments, the beam training module is further configured to, in the case of activating the central sparse linear uniform array, perform super-resolution angle scanning using multiple rainbow blocks based on the first beam splitting effect to obtain multiple candidate user angles;

[0319] Among them, the activated central sparse linear uniform array includes multiple antennas with equal antenna spacing; the interval between every two of the multiple rainbow blocks is less than zero, and the left edge of the first rainbow block among the multiple rainbow blocks covers the spatial angle -1, and the right edge of the last rainbow block covers the spatial angle 1.

[0320] In some embodiments, the beam training module is further configured to obtain the calibrated received power corresponding to each target subcarrier in the broadband; compare the magnitudes of the calibrated received powers to obtain the highest calibrated received power among the calibrated received powers; determine the second target subcarrier corresponding to the highest calibrated received power among the target subcarriers; and determine the candidate user angle covered by a single beam of the second target subcarrier among the candidate user angles as the actual user angle.

[0321] In some embodiments, the beam training module is further configured to, when all antennas of the activated ultra-large-scale array are activated, use the second beam splitting effect to control the beam of all subcarriers in the broadband to focus on a specific position in the actual user angle to obtain the user distance;

[0322] Among them, all antennas of the ultra-large-scale array are activated based on one pilot.

[0323] It should be noted that the specific implementation manners of the broadband near-field beam training device provided in the embodiments of the present application are basically the same as the specific embodiments of the above broadband near-field beam training method, and will not be elaborated herein.

[0324] The embodiments of the present application further provide a computer device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above broadband near-field beam training method is implemented. This computer device can be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0325] Please refer to Figure 18 , Figure 18 which schematically shows the hardware structure of a computer device in another embodiment. The computer device includes:

[0326] A processor 1801, which can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0327] The memory 1802 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1802 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1802 and are called by the processor 1801 to execute the broadband near-field beam training method of the embodiments of this application;

[0328] The input / output interface 1803 is used to implement information input and output;

[0329] The communication interface 1804 is used to implement communication interaction between this device and other devices. Communication can be achieved through a wired manner (such as USB, network cable, etc.) or through a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0330] The bus 1805 transmits information between various components of the device (such as the processor 1801, the memory 1802, the input / output interface 1803, and the communication interface 1804);

[0331] Among them, the processor 1801, the memory 1802, the input / output interface 1803, and the communication interface 1804 achieve communication connections with each other inside the device through the bus 1805.

[0332] The embodiments of this application also provide a computer-readable storage medium. This computer-readable storage medium stores a computer program, and when this computer program is executed by a processor, it implements the above-mentioned broadband near-field beam training method.

[0333] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory can optionally include memories remotely set relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and their combinations.

[0334] The embodiments of this application also provide a computer program product. This computer program product stores a computer program, and when this computer program is executed by a processor, it implements the above-mentioned broadband near-field beam training method.

[0335] The broadband near-field beam training method, broadband near-field beam training device, computer device, computer-readable storage medium, and computer program product provided by the embodiments of the present application utilize the dual beam splitting phenomena in the spatial domain and frequency domain based on a sparse array simultaneously, and can complete super-resolution beam alignment by using limited spectrum resources, greatly improving the accuracy of beam training. Moreover, the embodiments of the present application decouple the angle and distance estimation by using the activation method of the central sub-array, greatly reducing the pilot overhead required for beam training, and thus realizing a broadband near-field beam training method with super-resolution and extremely low overhead.

[0336] The embodiments described in the embodiments of the present application are for more clearly explaining the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0337] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown, or combine certain steps, or different steps.

[0338] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0339] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware, and their appropriate combinations.

[0340] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily need to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0341] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" may represent: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after. "At least one (one) of the following" or its similar expressions refer to any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0342] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the above division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in electrical, mechanical or other forms.

[0343] The units described above as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0344] In addition, each functional unit in each embodiment of this application can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0345] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs.

[0346] The preferred embodiments of the embodiments of this application have been described above with reference to the accompanying drawings, and thus do not limit the scope of the rights of the embodiments of this application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of this application shall be within the scope of the rights of the embodiments of this application.

Claims

1. A broadband near-field beam training method, characterized in that: The method comprises: Obtain system models for broadband very large-scale array communication systems; Determining a first beam splitting effect of a sparse array in a wideband in a spatial domain and a second beam splitting effect of a wideband in a frequency domain based on the system model; Super-resolution broadband near-field beam training is performed based on the first beam splitting effect and the second beam splitting effect.

2. The method according to claim 1, characterized in that The determining, based on the system model, a first beam splitting effect of a sparse array in a wideband in a spatial domain and a second beam splitting effect of a wideband in a frequency domain comprises: sparsely activating the system model to obtain a far-field channel model of a sparse linear uniform array; A first beam splitting effect of a sparse array in a wideband in a spatial domain and a second beam splitting effect of a wideband in a frequency domain are determined based on the far-field channel model.

3. The method according to claim 1, characterized in that The performing super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect includes: Performing super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles; Determine an actual user angle among the candidate user angles based on a first target subcarrier in a wideband; wherein a single beam of the first target subcarrier covers the candidate user angle; The actual user angle is estimated based on the second beam splitting effect to obtain a user distance.

4. The method according to claim 3, characterized in that The performing super-resolution angle estimation based on the first beam splitting effect to obtain multiple candidate user angles includes: In the case of activating the central sparse linear uniform array, based on the first beam splitting effect, multiple rainbow blocks are used to perform super-resolution angle scanning to obtain multiple candidate user angles; The activated central sparse linear uniform array includes multiple antennas with equal antenna spacing; the interval between every two rainbow blocks in the multiple rainbow blocks is less than zero, and the left edge of the first rainbow block in the multiple rainbow blocks covers a spatial angle of -1, and the right edge of the last rainbow block covers a spatial angle of 1.

5. The method according to claim 3, characterized in that: The determining the actual user angle among the candidate user angles based on the first target subcarrier in the broadband includes: Obtaining the calibrated received power corresponding to each target subcarrier in the broadband; Comparing the magnitudes of the calibrated received powers to obtain the highest calibrated received power among the calibrated received powers; Determine a second target subcarrier corresponding to the highest calibrated received power among the target subcarriers; The candidate user angle covered by the single beam of the second target subcarrier among the candidate user angles is determined as the actual user angle.

6. The method according to claim 3, characterized in that The performing distance estimation on the actual user angle based on the second beam splitting effect to obtain the user distance includes: When all antennas of the ultra-large-scale array are activated, the second beam splitting effect is used to control the beams of all subcarriers in the broadband to focus on a specific position in the actual user angle, so as to obtain the user distance; Here, all antennas of the ultra-large-scale array are activated based on one pilot.

7. A broadband near-field beam training device, characterized in that: The device comprises: An acquisition module, used for acquiring a system model of a broadband ultra-large-scale array communication system; A dual beam splitting effect determination module, used to determine a first beam splitting effect of a sparse array in a wideband in a spatial domain and a second beam splitting effect of a wideband in a frequency domain based on the system model; A beam training module is used to perform super-resolution broadband near-field beam training based on the first beam splitting effect and the second beam splitting effect.

8. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the broadband near-field beam training method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the broadband near-field beam training method according to any one of claims 1 to 6 is implemented.

10. A computer program product, wherein the computer program product stores a computer program, characterized in that: When the computer program is executed by a processor, the broadband near-field beam training method according to any one of claims 1 to 6 is implemented.

Images